Minor Project - 2

Integrated E-shopping application

MEMBER’S NAME ROLL NUMBER SAP ID BRANCH

DIVYAM VERMA R103218043 500067103 CSE-BAO

DEVANSH MESSON

R103218037 500067983 CSE-BAO

MAYANK RASTOGI R103218081 500069069 CSE-BAO

Under the guidance of

Dr.Amit Singh

Mentor Approval

CONTENTS

Introduction and Flow of project

Related work

Problem Statement and objective

Methodology

Progress and results

Conclusion, Future work and real-life application

References

Introduction

This project is representing an e-shopping application which can be used by customers. A particular category of

product is present in the database. Customers can efficiently select, safely purchase, and can get the selected

product really fast at his/her doorstep.

Extended Minor 1 project

Flow of the project

Data base containing 8000

unique mobile phones

Collaborative Filtering will suggest

top 10 best mobile phones on the

basis of likings of other users.

Applied sentiment analysis on the reviews

of these 10 best mobiles given by other

customers.It will further recommend the

best phone out of 10 by giving a score

value.

Customer will purchase the

product. An OTP will be received

by customer.

Encrypt OPT by RSA

algorithm for security.

Bi-directional Dijkstra algorithm

will compute the shortest path

from dispatch location to

customer’s location.

Customer looking for a

mobile in the database

by manually searching

Related work

Collaborative Filtering

Collaborative filtering is a method that can filter out things that can that a user may like based on response by other

users. It works via looking through a huge group of individuals. and finding a smaller set of users with tastes similar

to a particular user. It looks at the things they like and combines them to create a ran ked list of suggestions [1].

Sentimental Analysis

Sentiment analysis (or opinion mining) is a natural language processing technique decide if the information is

positive, negative, or unbiased. Sentiment analysis is frequently performed on text-based information to help

organizations and item sentiment in user’s fee dback, and comprehend user’s needs [2].

Encryption-Decryption

Encryption is the process when plaintext is changed over into obviously arbitrary secret message for security reason refer red to as

ciphertext [3].

Decryption is the process of ch anging the ciphertext to the plaintext that process is known as decryption [4].

Bi-directional Dijkstra

Bi-directional method runs the Dijkstra algorithm from the source vertex and destination vertex at the same time

and the algorithm ends when a vertex is processed by forward binary heap and backward binary heap [5].

Problem Statement

In this pandemic, people are forced to stay at home as it becomes risky to go outside

and do shopping in crowded areas. Therefore, E-shopping becomes an easy and a safer

alternative for shopping goods, which also makes the experience more helpful and

fulfilling for a buyer than conventional shopping.

• Recommend the best and customized product for the customer on the basis of customer’s liking of brand of

product, customer’s budget and reviews given by different customers who had just purchased that specific

item.

• Secured online transactions and conceal buyer's financial information using encryption-decryption

• Suggest the delivery boy, the shortest possible path as soon as possible in the real road network by

applying time efficient shortest path algorithms .

Objective

Methodology

For Recommendation System-

2 methods are being proposed:

• Collaborative Filtering

• Sentimental Analysis

For Encryption-Decryption -

2 methods are being proposed:

• Vi-genre Cipher (Less strong)

• RSA algorithm (More strong)

For solving Time Efficient Shortest Path-

4 solving optimization is being proposed:

• Brute-force method

• Fibonacci heap implementation

• Binary heap implementation

• Bi-directional Dijkstra algorithm (Giving Least execution time)

Recommendation Collaborative Filtering Sentiment analysis

Use - Case Diagram

Represents the flow of our whole project

Algorithm of Collaborative filtering

What it does in project:

It will recommend top 10 best mobile phones out of 8000 phones.

Algorithm of Sentiment analysis

What it does in project:

From top best 10 mobile phones, It

will shortlist the best phone out of 10

phones.

Formula of Naïve Bayes algorithm

Optimizing

Dijkstra algorithm

Pseudocode

Bidirectional

Dijkstra algorithm

What it does in project:

It will help in delivering the product

faster to customer’s place.

Progress after Mid-Sem Presentation

 Deployed recommendation system consisting of collaborative filtering as a web application on local host using Javascript, Python, HTML,

CSS

 Integrated the 3 modules of project in C++ language.

Home page

Results

For

Sentiment Analysis using Naïve Bayes classification:

Trained 4000 reviews (80% of 5000 reviews)

Tested 1000 reviews (20% of 5000 reviews)

Achieved a validation accuracy of 86%.

Sentiment Analysis and Collaborative Filtering

Collaborative Filtering (User-based):

Recommended top 10 best mobile phones out of

8000 mobile phones for a particular customer.

Results

For

Encryption

RSA encryption algorithm : (BEST)

Strongly encrypted the one time password.

Suppose OTP = “devanshmesson”

Vigenere Cipher encryption algorithm :

Encrypted the one time password.

Suppose OTP = “DEVANSHMESSON”

More strong :

Less strong :

To hack it : try 26^13 permutations!

To hack it : try more than 26^13

permutations because of special characters.

Results

For

Proposed solutions of Dijkstra algorithm

37 seconds

3 milliseconds

Increasing density of graph

Why Bi-directional Dijkstra is best?

Future work :

• Will use TFIDF vectorizer and word2vec for improved accuracy.

• Will Implement content based filtering for improved recommendation.

• Will try to implement AES algorithm for improved encryption.

• Will enhance the web application by including all modules of this project

Conclusion

• In the case of recommendation system both collaborative filtering and sentiment analysis worked together to recommend the best

phone/s out of 8000 mobile phones on the basis of 40,000 user reviews and ratings.

• In the case of encryption- decryption, after comparing the security level of RSA and Vigenère cipher, The RSA performs better than

Vigenère cipher because it provides 2 key encryption which is based on prime numbers and as well as to decrypt the ciphertext

without keys we required more than 26^13 permutations as it uses special symbols with numbers and alphabets.

• In the case of shortest path algorithm, after closely comparing the execution time of four algorithms on various real-world test cases,

Bidirectional Dijkstra algorithm comes out to be the most efficient algorithm out of the four algorithms on the basis of time complexity.

Real life application

This project is very beneficial for current pandemic situation. Online ordering of goods is 10 times safer than visiting a crowdy

place. Moreover, Bi-directional Dijkstra algorithm provides efficiency to suggest the shortest path for speedy delivery.

Encryption algorithm provides safety while transaction. Moreover, Recommendation is very beneficial for effective shopping.

References

1) G. Linden, B. Smith and J. York, "Amazon.com recommendations: ite m-to-item collaborative filtering," in IEEE Internet Computing,

vol. 7, no. 1, pp. 76-80, Jan.-Feb. 2003, doi: 10.1109/MIC.2003.1167344.

2) B. K. Bhavitha, A. P. Rodrigues and N. N. Chiplunkar, "Comparative study of machine learning techniques in sentiment al analysis,"

2017 International Conference on Inventive Communication and Computational Technologies (ICICCT), Coimbatore, 2017, pp. 216-

221, doi: 10.1109/ICICCT.2017.7975191.

3) D. N. Wu, Q. Q. Gan and X. M. Wang, "Verifiable Public Key Enc ryption With Keyword Search Based on H omomorphic

Encryption in Multi-User Setting," in IEEE Access, vol. 6, pp. 42445-42453, 2018, doi: 10.1109/ACCESS .2018.2861424.

4) https://www.geeksforgeeks.org/vigenere-cipher/

5) Rivest, R.; Shamir, A.; Adleman, L. (February 1978). "A Method for Obtaining Digital Signatures and Public-Key

6) E.W Dijkstra, “A note on two problems in connection with graphs” Numerische Mathematik, vol. 1 , pp. 269–271, 1959.

7) T.H. Cormen, C.E. Leiserson, R.L. Rivest and C. Stein, "Dijkstra's al gorithm ". Introduction to Algorithms, 2nd ed., MIT Press and

McGraw–Hill. pp. 595–601, 2001, ISBN 0-262-03293-7.

8) T. Porter and I. Simo n, "Random insertion into a priority queue structure". IEEE Transactions on Software Engineerin g. SE-1 (3),

292–298, 1975, doi:10.1109/TSE.1975.6312854

9) T.B. de Silveira, E. M. Duque, S.J.F. Guimarães, H. T. Marques-Neto and H. C. de Freitas, “Proposal of Fibonacci heap in the

Dijkstra algorithm for low-power ad-hoc mobile transmissions”. IEEE Latin America Transactions, vol. 18, issue 3, pp. 623-630, 2020,

doi:10.1109/TLA.2020.9082735

10) L. Chen, J. Zhou, J. Li and Y. Chen, “Bidirectional Dijkstra algorithm for best-routing of urban traffic network”. Proceedings SP IE,

vol. 6754, Geoinformatics 2007: Geospatial Information Technology and Applications, 675418 , 2007,

https://doi.org/10.1117/12.764679.